Using Mathematical Morphology to Detect the Imperfections of the Printed Circuit Boards

Abstract

Image processing is a form of signal processing(analysis, interpretation, and manipulation of signals) forwhich the input is an image, such as photographs; the outputof image processing can be either an image or a set ofcharacteristics or parameters related to the image. Typicaloperations of image processing are geometrictransformations such as enlargement, reduction androtation, color corrections such as brightness and contrastadjustments, quantization, or conversion to a different colorspace, interpolation and recovery of a full image from a rawimage format, image editing, image differencing, imagestabilization, image segmentation and other. Applications ofthese operations are encountered in computer vision, facedetection, feature detection, medical image processing,microscope image processing, morphological imageprocessing, remote sensing and many other disciplines.Morphological image processing is a collection of techniquesfor digital image processing based on mathematicalmorphology. By combining these morphological operatorswe can obtain algorithms for many image processing tasks,such as feature detection, image segmentation, imagesharpening, image filtering, and granulometry. Likewise,using this technology we can detect some imperfections of theprinted circuits, imperfections that allow us to repair it,opportunely. So in this way, we can eliminate some costs inthe technological process.The paper presents an introduction in mathematicalmorphology, describes “hit and miss transform”, dilationand erosion and, after all, we present the usage of two ofthese operations in printed circuit error detection.

Keywords

• morphology
• image
• erosion
• dilation
• opening
• closing
• pixel